Research · Ocean Dynamics from Space

SWOT

The Surface Water and Ocean Topography (SWOT) mission—led by NASA and CNES, with UK and Canadian partners—launched on 16 December 2022. It measures water elevation across oceans, lakes, rivers, and coasts at unprecedented resolution.

Our laboratory focuses on SWOT’s ocean dataset: classical physical oceanography for small-scale dynamics, together with modern tools such as machine learning for feature identification.

Selected publications

Archer, M., Wang, J., Klein, P. et al. Wide-swath satellite altimetry unveils global submesoscale ocean dynamics. Nature 640, 691–696 (2025)
A short-form demonstration of SWOT’s capability for submesoscale ocean dynamics.
Wang et al., On the development of SWOT in situ calibration/validation for short-wavelength ocean topography (2022), J. Atmos. Oceanic Technol.
Foundation for the post-launch cal/val campaign (March–July 2023, SWOT fast-repeat phase).
d’Ovidio et al., Frontiers in fine-scale in situ studies: Opportunities during the SWOT fast sampling phase (2019), Front. Mar. Sci.
Community expectations for SWOT before launch.

Ocean eddies

Eddies drive over 80% of the ocean’s kinetic energy and shape the transport of heat, nutrients, and tracers. Mesoscale eddies (hundreds of kilometers) are well observed by altimetry; submesoscale eddies (kilometers to tens of kilometers) remain less understood yet are vital for vertical exchange.

We measure and reconstruct their horizontal and vertical structure from incomplete datasets, combining satellite and in situ observations with modeling and theory.

Selected publications

Wang J., Flierl G.R., LaCasce J.H., McClean J.L., Mahadevan A. Reconstructing the ocean’s interior from surface data (2013), J. Phys. Oceanogr.
Decouples interior and surface dynamics using SSH and SST to reconstruct vertical eddy structure.

Machine learning for ocean science

With collaborators in machine learning, we apply ML to oceanographic data—especially SWOT—for feature identification, data assimilation, and predictive modeling.

Selected publications

Manucharyan, G. E., Martin, S. A., Balwada, D., and Wang, J.: Deep Learning to Infer Ocean Dynamics, Annu. Rev. Mar. Sci., 19, in press, 2026
Reviews deep learning for inferring ocean dynamics from incomplete observations, including interpolation, neural data assimilation, forecasting, and hybrid eddy parameterizations.
Goh, E., Yepremyan, A., Wang, J., and Wilson, B.: MAESSTRO: Masked Autoencoders for Sea Surface Temperature Reconstruction under Occlusion, Ocean Sci., 20, 1309–1323, 2024
Adapts masked autoencoders to reconstruct missing satellite imagery under cloud occlusion.

Geophysical fluid dynamics

GFD is the cornerstone of physical oceanography: understanding observed phenomena and improving models of circulation and climate. Flows are nonlinear and chaotic; we simplify, solve, and simulate to reveal mechanisms of energy transfer and jet formation.